Low-elevation endemic Rhododendrons in China are highly vulnerable to climate and land use change

(1)

Ecological Indicators 126 (2021) 107699

Available online 16 April 2021

Low-elevation endemic Rhododendrons in China are highly vulnerable to

climate and land use change

Fangyuan Yu

a

_{, Zhifeng Wu}

a,*

_{, Jian Shen}

b

_{, Jihong Huang}

c

_{, Thomas A. Groen}

d

_,

Andrew K. Skidmore

d,e

_{, Keping Ma}

f

_{, Tiejun Wang}

d

a_{School of Geographical Sciences and Remote Sensing, Guangzhou University, Guangzhou 510006, China}

b_{Institute of Agricultural Resources and Environment, Guangdong Academy of Agricultural Sciences, Guangzhou 510640, China}

c_{Key Laboratory of Biodiversity Conservation of National Forestry and Grassland Administration, Institute of Forest Ecology, Environmental and Protection, Chinese} Academy of Forestry, Beijing 100091, China

d_{Department of Natural Resources, Faculty of Geo-Information Science and Earth Observation, University of Twente, P.O. Box 217, 7500 AE Enschede, the Netherlands} e_{Department of Environmental Science, Macquarie University, NSW 2109, Australia}

f_{State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China}

A R T I C L E I N F O Keywords: Indicator Biodiversity Conservation Weighted endemism Range shift Protected areas A B S T R A C T

The combination of climate change and land use change may have profound effects on terrestrial biodiversity in more significant ways than either has separately. However, most studies focus largely on the climate change impacts, which hampers our ability to develop appropriate conservation strategies in a dramatically changing world. Here, we predict the distributions of 191 Chinese endemic Rhododendron species under future climate and land use change, combining two dispersal constraint scenarios by using a species distribution model. We then assess the vulnerability and extinction risk of these species and identify areas at risk of highest species loss. We find that 52% of the species are predicted to expand and shift their geographic ranges, typically to the northwest and north. The remaining 48% of species are predicted to contract in geographic ranges under the ’perfect- dispersal’ scenario. And only 7% of Rhododendron are predicted to keep ’still’, while the rest of species shrank with varying degree under the ’no-dispersal’ scenario. Species lost particularly at lower elevations, and we also identify four regions at particularly high risk from the impacts of climate and land use change, namely the parallel ridge-and-valley areas of eastern Sichuan, southeastern Tibet, western and eastern Yunnan, southern Shaanxi, plus scattered areas in Guangdong, Hainan and Taiwan. We conclude that Chinese endemic Rhodo-dendron species at lower elevations are highly vulnerable to climate and land use change, facing an elevated risk of extinction under varying scenarios. These species therefore call for more attention and protection. We high-light the critical role of endemic Rhododendron species as good indicators for measuring, evaluating and un-derstanding the effectiveness of our biodiversity conservation efforts. Our work provides insight into the status, trends and threats regarding endemic Rhododendron species, identifying risks and prioritizing conservation in a rapidly changing world.

1. Introduction

Biodiversity continues to decline across the globe under the effect of multiple anthropogenic pressures (Osborn et al., 2019). Climate change and land use change are two major threats to global biodiversity. The cumulative or synergistic effect of climate change and land use change is believed to have an even greater impact on biodiversity over the coming

century (Brook et al., 2008; Sala et al., 2000). However, biodiversity responses to changes in climate and land use certainly appear hetero-geneous, varying among species and geographical regions (Osborn et al., 2019). As regional land use change may either exacerbate or alleviate climatic impacts on biodiversity, it is important to integrate both factors to better understand potential impacts of future climate and land use changes on biodiversity (Oliver and Morecroft, 2014). However, most * Corresponding author at: School of Geographical Sciences and Remote Sensing, No. 230 Wai Huan Xi Road, Guangzhou Higher Education Mega Centre, Guangzhou 510006, China.

E-mail addresses: yfy@gzhu.edu.cn (F. Yu), zfwu@gzhu.edu.cn (Z. Wu), shenjian@gdaas.cn (J. Shen), t.a.groen@utwente.nl (T.A. Groen), a.k.skidmore@utwente. nl (A.K. Skidmore), kpma@ibcas.ac.cn (K. Ma), t.wang@utwente.nl (T. Wang).

Contents lists available at ScienceDirect

Ecological Indicators

journal homepage: www.elsevier.com/locate/ecolind

https://doi.org/10.1016/j.ecolind.2021.107699

(2)

Ecological Indicators 126 (2021) 107699 previous studies tend to focus on the impacts of climate change alone,

with much less consideration given to the synergistic effects of climate and land use change, thus potentially hampering our ability to develop appropriate conservation strategies (Sirami et al., 2017).

When conservation resources are limited, identifying priority areas for the creation of protected areas is often considered to be the essential strategy to halt the decline in biodiversity (Xu et al., 2018). Areas not included in existing protected zones, but at high risk of losing species, should be considered conservation priority areas (Chen et al., 2017). However, due to climate and land use change, species may also shift their distribution range and move out of current protected areas (Ala-gador et al., 2014). It is therefore necessary to assess the suitability of protected areas for species in response to future environmental change (Chen et al., 2017). Consequently, a better understanding of protected areas and biodiversity dynamics under future climate and land use change is needed (Thomas et al., 2004; Thuiller et al., 2011).

Using focal species to determine the status of biodiversity is a po-tential tool for developing and evaluating conservation strategies (Ye et al., 2018). The genus Rhododendron is composed of more than 1000 species globally, of which nearly 70% is classified as vulnerable, threatened, endangered or critically endangered (Gibbs et al., 2011). Of the approximately 571 Rhododendron species present in China, 422 are endemic (Ma et al., 2014; Wu et al., 2005). It is worth to note that

Rhododendrons form a major component of the montane ecosystem in

the Himalayan subalpine and alpine zone (Kumar, 2012). The whole Himalayan region, including the Tibetan Plateau is regarded as one of the world’s most critical centers of biodiversity (Myers et al., 2000). The Himalayan region has displayed a warming trend greater than the global average of 0.74 ◦_{C (IPCC, 2013). As well, the ever-increasing pressures} on the land to service the needs of an expanding population are placing some Rhododendron species at risk of extinction (Ma et al., 2014). As the genus Rhododendron has an extremely complex taxonomic structure and displays large morphological variation, it is an excellent candidate for conservation planning in response to future climate and land use change (MacKay & Gardiner, 2016).

The aim of this study is to understand the synergistic impacts of future climate and land use change on Chinese endemic Rhododendron species. Specifically, we set out to (1) predict potentially suitable habitat for Chinese endemic Rhododendron species as well as their vulnerability and risk of extinction; (2) estimate the change in diversity patterns of Chinese endemic Rhododendron species; (3) identify priority areas for the conservation of Chinese endemic Rhododendron species.

2. Material and methods

2.1. Chinese endemic Rhododendron species

We collected the distribution data of Rhododendron species from seven Chinese herbaria and botanical museums (for more details, see Yu et al., 2015) at county level, and the latest Flora of China (Wu et al., 2005). After removing ambiguous records, our database included 406

Rhododendron species out of the 571 species occurring in China. In total,

we obtained 13,126 geo-referenced records, with each record having a spatial uncertainty of less than 1 km. Given the focus on endemic species in this study, and the aspiration to assure robustness of the results of the species distribution modelling, only endemic species with more than 5 occurrences were retained. On the one hand, we endeavored to identify priority conservation areas under future environmental change, including as many Rhododendron species as possible to form a compre-hensive overview. On the other hand, there should also be a focus on selecting conservation priority areas for rare Rhododendron species either occupying a small range or enduring low abundance (Synge et al., 1981), while being careful not to neglect the vast majority of species (ter Steege et al., 2013; van Proosdij et al., 2016). Eventually, we narrowed the dataset down to 191 endemic species, with 9270 records in total. The geographic coordinates of each record were projected onto an Albers

Equal-area Conic Conformal coordinate system to avoid the latitudinal bias of geographic coordinate systems.

2.2. Environmental variables

We initially downloaded 19 biologically relevant climatic variables from the WorldClim database (http://www.worldclim.org/) with a spatial resolution of 30 arc second (approximately 1 km at the equator). Based on collinear relationships (Pearson r < 0.7) as well as the physi-ology and life history requirements of Rhododendron, six climatic vari-ables comprising isothermality (Bio3), temperature seasonality (Bio4), minimum temperature of coldest month (Bio6), annual precipitation (Bio12), precipitation of wettest month (Bio13), and precipitation of driest month (Bio14) were ultimately selected. Values for recent (the average for 1950–2000) and for the 2070s (the average for 2061–2080) climatic conditions were derived. Because variability in general circu-lation model (GCM) projections forms a major source of uncertainty in predictive species modelling (Sommer et al., 2010; Thuiller et al., 2005), we selected to employ three widely-used general circulation models (GCMs), namely the Beijing Climate Center Climate System Model (BCC- CSM1-1), the Community Climate System Model (CCSM4), and the Hadley Global Environment Model 2 - Earth System (HadGEM2-ES), to project future species distributions. For each GCM, four Representative Concentration Pathway (RCP) scenarios (RCP2.6, RCP4.5, RCP6.0, RCP8.5) were considered, which each make widely differing assump-tions about possible future socio- economic pathways (van Vuuren et al., 2011). Of these four RCPs, we selected RCP2.6 and RCP8.5, which represent the most “benign” scenario (i.e., a likely increase of 0.3–1.7 ◦_C for ca. 2081–2100) and the most extreme scenario (i.e., a likely increase of 2.6–4.8 ◦_{C for ca. 2081–2100), respectively, for bracketing the likely} futures (van Vuuren et al., 2011). In addition, we included topographic and edaphic variables, represented by elevation with a resolution of 90 m (http://www.srtm.csi.cgiar.org) and pH of the top 30 cm of soil (Hengl et al., 2014, https://www.isric.org/explore/soilgrids), respec-tively, to model species distribution. Given that topography and soil condition hardly change within short periods of time (Merow et al., 2014), we kept these two variables consistent for both the current and future conditions.

We derived current (the period around 2010) and future (the period around 2070) global land use data at a 30 m resolution from the Finer Resolution Observation and Monitoring-Global Land Cover (FROM-GLC (Li et al., 2016), available at http://data.ess.tsinghua.edu.cn/), which includes 10 general (cropland, forest, grassland, shrubland, wetland, water, tundra, impervious, bare land, and snow/ice) and 28 detailed land use types. The future land use projections were set up based on the set of global change scenarios. For consistency with the climatic data, we used the land use datasets under RCP2.6 and RCP8.5 as representative for the future land use scenarios. We resampled all land use data using a ‘majority’ resampling method to the same spatial resolution (1 km × 1 km) of the other environmental variables in ArcGIS 10.2 (ESRI, 2011).

2.3. Species distribution modeling

A number of species distribution models as well as the ensemble modeling approach have been proposed to model future species distri-bution and diversity (Franklin, 2010a). Since we had presence-only data of Rhododendron species, we built individual models for each species using maximum entropy (MaxEnt) modeling (Phillips et al., 2006). The MaxEnt model is optimized to work with presence-only data and is known to also produce robust results with small sample sizes, e.g., 5 occurrences (Aguirre-Gutierrez et al., 2015; Pearson et al., 2007; Rana et al., 2021; van Proosdij et al., 2016). We kept the maximum number of background points for sampling at 10,000, and set the maximum itera-tion for each run to 1000. We ran a 10-fold cross validaitera-tion for the potential distribution. The relative importance of each of the included predictor variables was assessed using a jackknife test, which is a F. Yu et al.

(3)

Ecological Indicators 126 (2021) 107699

standard output of MaxEnt (Phillips et al., 2006).

Model performance was evaluated by calculating both the area under the receiver operating characteristic curve (AUC), a threshold- independent measure of model accuracy (Swets, 1988), and the true skill statistic (TSS), a threshold-dependent measure of model accuracy (Allouche et al., 2006). AUC has a range from 0.5 to 1, where 1 indicates a perfect fit for the model, over 0.7 indicates a good fit, and around 0.5 suggests randomness (Swets, 1988). TSS ranges from − 1 to 1, where values of 1 indicate perfect predictions, 0 indicate random predictions, and − 1 indicate perfect inverse predictions. TSS considers the measures of both sensitivity and specificity so that both omission and commission errors are accounted for.

The default logistic output of MaxEnt is a continuous variable ranging from 0 to 1, where high values indicate higher relative suit-ability. We created binary distribution maps of suitable and unsuitable habitats from the continuous probability predictions of each species by using the threshold where TSS was maximal (Liu et al., 2013). The MaxTSS threshold can minimize the mean of the error rate and has been widely used in species distribution modeling (SDMs, Liu et al., 2013). We averaged the final binary result for each species from 10 times’ predictions across the three GCMs.

To investigate how future land use change would affect species dis-tributions, we repeated the SDMs with two sets: (1) climate change only (CC), using the future climate conditions but fixing the land use variable to the observed 2010 land use data, (2) both climate and land use change (CCLC): using both the future climate and land use conditions of the 2070s. The relative effect of future land use change was then examined by comparing the results from the two different scenarios.

2.4. Dispersal scenarios

Given the lack of accurate data to calibrate the dispersal behavior of each species, we estimated species shift under two scenarios (Feeley and Silman, 2010): (1) ’perfect-dispersal’ scenario, assuming Rhododendron are able to occupy all habitat areas that is suitable under future envi-ronment change. (2) ’no-dispersal’ scenario, assuming Rhododendron are incapable to migrate because of the limited dispersal abilities of them-selves or other limiting factors of environment, and only occupy the habitat which is suitable under both current and future environment.

2.5. Quantifying species responses

To quantify species’ response to future climate and land use change, we estimated the change in distribution range for each species, including range contraction, range persistence, and range expansion, by comparing the value of each grid cell for each species’ current and future distribution range. We further calculated the relative change in the total area of suitable habitat (CSH) for estimating species vulnerability with

CSH =Af− Ac Ac *100

where Af and Ac are suitable area under future and current conditions, respectively. Additionally, the changing distance and direction of spe-cies’ range shifts were estimated based on range centroids of current and future binary SDM prediction distributions. We carried out these two steps with SDMtoolbox 2.0 (Brown et al., 2017).

Fig. 1. Relative change in the total area of suitable habitat (CSH, ranging from − 100% to >100%) to climate change only and to combined climate and land use

change for RCP2.6 and RCP8.5 for the 2070s (2071–2100) with ’perfect-dispersal’ scenario.

(4)

2.6. Measure of diversity change

We used two metrics to evaluate changes in species diversity: species richness (i.e. the number of unique species in a geographical unit) and weighted endemism (i.e. proportional to the inverse of a species’ range; Laffan & Crisp, 2003). To determine the change in species richness, we stacked the individual binary species distribution maps (Distler et al., 2015), producing the total number of species in each geographical unit of 10 × 10 km. We chose this grid cell size of 10 × 10 km to measure

Rhododendron diversity for two reasons. On the one hand, we aspired to

provide as much detail as possible regarding the spatial patterns of

Rhododendron diversity, but on the other hand, a practical unit for

conservation planning was required in order to efficiently allocate scarce resources based on previous study (Yu et al., 2017). Next, we calculated the richness change as the difference between the current number of species and the future number of species. Negative numbers indicate losses and positive numbers gain in species occurrences.

To determine the change in weighted endemism, weighted ende-mism was calculated according to the same method as outlined by Herkt et al. (2016):

Weighted endemism =∑

t∈T 1

Rt

where T is the total number of species found in the study area, t is a species in T, and Rt is the range size of species t. Endemism change was then estimated from the difference between future and current weighted endemism cell by cell similarly to the way this was achieved for species richness.

In addition, we explored the relationship between diversity change

and elevation by calculating the distribution of species loss and species gain with regards to elevation. Specifically, we extracted elevation from change in both species richness and weighted endemism, and plotted this change against elevation. We also calculated the degree of change in species richness and weighted endemism as elevation increased.

2.7. Conservation gap analysis

Conservation gaps were identified by overlapping the modelled changes in species richness and weighted endemism with the existing protected area system (i.e., 2139 Chinese nature reserves). The top 1% area of loss in species richness or weighted endemism was intersected with the Chinese nature reserves. The priority areas for protecting

Rhododendron species were the areas losing most Rhododendron species

or weighted endemism, but were not yet covered by a protected area. All spatial analyses were conducted using R 3.3.2 (R Core Team, 2015) and ArcGIS 10.2 (ESRI, 2011).

3. Results

3.1. Changes in suitable habitat area

We obtained vigorous species distribution models for the 191 Chi-nese endemic Rhododendron species with AUC values ranging from 0.70 to 0.99, and TSS varying from 0.50 to 0.99 under RCP2.6. RCP8.5 showed similar results, with AUC ranging from 0.50 to 0.99, and TSS varying from 0.48 to 0.99 (Fig. S1).

Under the ’perfect-dispersal’ scenario, in general, we predicted an average of 48% of species would lose habitat, while 52% of species

Fig. 2. Relative change in the total area of suitable habitat (CSH, ranging from − 100% to 100%) to climate change only and to combined climate and land use

change for RCP2.6 and RCP8.5 for the 2070s (2071–2100) with ’no-dispersal’ scenario.

(5)

would gain habitat under different scenarios. Under the RCP2.6 sce-nario, 88 and 94 species were predicted to experience a contraction of their suitable habitats under the climate change and under the combined climate and land use change scenarios, respectively; While 103 and 97 species, were predicted to expand their suitable habitats under the two scenarios, respectively (Fig. 1a and b). By contrast, with the RCP8.5 scenario, we predicted nearly equal numbers would lose (92 and 96 species, respectively) and gain (99 and 95 species, respectively) suitable habitats under the climate change and the combined climate and land use change scenarios (Fig. 1c and 1d).

If Rhododendron are incapable to migrate with future environment, they generally were predicted to lose their habitats under varying sce-narios (Fig. 2). Under the ’no-dispersal’ scenario, about 13

Rhododendron (7% of total) species would keep ’still’, and the remaining

species were projected to lose their habitats under climate and land use change of RCP2.6 and RCP8.5. Meanwhile, about 3 Rhododendron spe-cies might go extinction because of completely loss of habitat. Consistent with the ’perfect-dispersal’ scenario, species were inclined to lose more habitats under the combined climate and land use scenarios and RCP 8.5 then the climate-only and RCP2.6 scenario.

3.2. Distribution shifts

Under the ’perfect-dispersal’ scenario, on the whole, most

Rhodo-dendron species were projected to shift in a north-westerly direction,

under both RCP2.6 and RCP8.5, with the shifting speed being relatively

Fig. 3. Centroid changes in direction and distance of species range between current and projected distributions under a climate change only scenario [(a) and (c)]

and a combined climate and land use change scenario [(b) and (d)] of RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] for the 2070 s (2071–2100) with ’perfect- dispersal’ scenario. The centroid change of each species was assigned to one of eight directions. Different colours in the figures represent the distance of species range shifts, while the length of each colour bar indicates the number of species at that projected range shift distance interval.

(6)

slower under climate change alone than under combined climate and land use change, as well as under RCP2.6 than under RCP8.5. Under RCP 2.6, a total of 71 (37%) and 80 (42%) species were projected to shift northwestwards over a distance of 78 km under climate change alone and 67 km under combined climate and land use change, while, 48 (25%) and 35 (18%) species would shift in a westerly direction, over a distance of 87 and 107 km, respectively (Fig. 3a and b).

By contrast, more species would shift northwestwards and north-wards under the RCP8.5 scenario, with 87 (46%) and 89 (47%) species moving in a northwestward direction, respectively, when considering climate change alone and combined climate and land use change. The shifting distances were also relatively longer than for RCP2.6, with

distances of 83 and 82 km, respectively, covered in a northwesterly di-rection. As well, 45 (24%) and 40 (21%) species were projected to move northwards, over distances of 35 and 29 km (Fig. 3c and d).

Under the ’no-dispersal’ scenario, the shifting direction were generally consistent with the ’perfect-dispersal’ scenario, typically to the northwest, north and west direction. However, the shifting distance were only less than 5 km (varying from 4 km to 7 km) under various climate and land use change of RCP2.6 and RCP8.5 scenarios (Fig. 4).

3.3. Spatial changes of species richness and weighted endemism

Under the ’perfect-dispersal’ scenario, the change map of species

Fig. 4. Centroid changes in direction and distance of species range between current and projected distributions under a climate change only scenario [(a) and (c)]

and a combined climate and land use change scenario [(b) and (d)] of RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] for the 2070 s (2071–2100) with ’no-dispersal’ scenario. The meaning of colour and bar are explained in Fig. 3.

(7)

richness showed that the central and southeastern part of China become less suitable, while the southwestern part becomes more suitable for a number of endemic Rhododendron species under the different future climate and land use change scenarios of RCP2.6 (Fig. 5a and b) and RCP8.5 (Fig. 5c and d). Specifically, areas that would lose species richness were mainly located between eastern Sichuan and western Chongqing, southern Shaanxi, southeastern Guizhou, and the central mountainous areas of Taiwan. Southeastern Tibet and western Sichuan were the regions which would gain most (top 10%) species under the two scenarios. When comparing RCP2.6 to RCP8.5, some areas would lose species according to RCP2.6, that would expand under RCP8.5, especially in the regions of eastern Sichuan and northwestern Yunnan. Overall, the joint effect of climate plus land use change on Rhododendron species richness was stronger than the effect of climate change alone.

It is worth noting that Sichuan province showed a contrasting pattern: while the eastern part would lose many Rhododendron species,

the western part would gain more species than the other provinces in China. The negative effect of land use change on the distribution of

Rhododendron in Sichuan was prominent.

Under the ’no-dispersal’ scenario, areas (top 5%) of southern Shaanxi, eastern Sichuan, western Chongqing, and northwestern Yunnan would lose species richness severely, and the combined climate and land use scenarios and RCP8.5 then the climate-only and RCP2.6 scenario, which were in line with the ’perfect-dispersal’ scenario (Fig. 6). In addition, under either ’perfect-dispersal’ or ’no-dispersal’ condition, we found that species loss occurred at relatively low eleva-tions (<2000 m), while species gain occurred at higher elevaeleva-tions, especially between 3000 and 5000 m for both the climate change and the climate and land use change scenarios of RCP2.6 (Fig. 7a and b) as well as RCP8.5 (Fig. 7c and d) of ’perfect-dispersal’ scenario. The mean elevation of areas, which would see a decrease in species richness, was 970 m (varying from 50 to 1200 m) and 880 m (varying from 600 to

Fig. 5. Spatial pattern of changes in Chinese endemic Rhododendron species richness under climate change alone [(a) and (c)] and under combined climate and land

use change [(b) and (d)] for RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’perfect-dispersal’ scenario in China for the 2070 s (2071–2100). F. Yu et al.

(8)

2400 m) for the climate change scenario and the climate and land use change scenario of RCP2.6, respectively. The mean elevation of regional lost and gained species under climate and land use change was about 100 m higher for RCP8.5 than for RCP2.6. Under the ’no-dispersal’ scenario, species lost occurred at the elevation of about 1400 m under climate and land use change for both RCP2.6 (Fig. 8a and b) and RCP8.5 (Fig. 8c and d).

Under the ’perfect-dispersal’ scenario, the projected change of weighted endemism was consistent with the change pattern of species richness (Fig. 9). The variation in weighted endemism also mainly occurred in the southern part of China. The decline of weighted ende-mism was most remarkable in northern Yunnan, southeastern Tibet, along the border of Sichuan and Shaanxi. The change in weighted endemism was also more striking when considering both climate and land use change than when considering climate change alone.

The projected change of weighted endemism for RCP2.6 and RCP8.5

of ’no-dispersal’ scenario was similar with the ’perfect-dispersal’ to a large extent (Fig. 10). However, more endemic species lost occurred at the southwest mountains, including Mt. Daba (i.e., the border of Sichuan and Shaanxi), Mt. Wandou and Mt. Dalou (i.e., the border of Sichuan and Chongqing).

Interestingly, under the ’perfect-dispersal’ scenario, loss of weighed endemism occurred over a broad range (from 100 m to 3000 m), while gain of weighed endemism mainly occurred at higher elevation (1000 ~ 2500 m and 3500 ~ 6000 m) under both the climate and the climate and land use change scenario of RCP2.6 (Fig. 11a and b) and RCP8.5 (Fig. 11c and d). Noteworthy is that the region’s lost weighted ende-mism occurs at a 100 m higher elevation under RCP8.5 than under RCP2.6.

Compared to the ’perfect-dispersal’ scenario, the variation of pro-jected change of weighted endemism along elevation under the ’no- dispersal’ scenario were slighter, but showed a similar trend. Loss of

Fig. 6. Spatial pattern of changes in Chinese endemic Rhododendron species richness under climate change alone [(a) and (c)] and under combined climate and land

use change [(b) and (d)] for RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’no-dispersal’ scenario in China for the 2070 s (2071–2100). F. Yu et al.

(9)

weighed endemism occurred at lower elevation (about 1100 m ~ 1300 m), while gain of weighed endemism mainly occurred at higher eleva-tion (about 2700 m ~ 3000 m) under both the climate and the climate and land use change scenario of RCP2.6 (Fig. 12a and b) and RCP8.5 (Fig. 12c and d).

3.4. Priority areas for conservation

Areas shown to be at high risk of losing Rhododendron species under various scenarios were only partly (ranging from 0.4% to 12.5%) covered by the current 2139 Chinese nature reserves (Table 1, Fig. 13 & Fig. 14).

Specifically, under both ’perfect-dispersal’ and ’no-dispersal’ sce-nario, an area of 16666 km2, including southeastern Tibet, the border of

Sichuan and Shaanxi (Mt. Daba), and southern Taiwan, would be at high risk of decline in species richness under the climate change only of RCP2.6 scenario. While an area of 21620 km2_{, mainly at Dazhou,} Guangan (Sichuan), Kunming and Mengzi (Yunnan), was not covered by any existing protected areas under the climate and land use change of RCP2.6 scenario. The priority areas not covered by the existing nature reserves were even larger under RCP8.5, involving areas at Motuo (southeastern Tibet), Dali and Tengchong (western Yunnan), from Guangyuan to Xichang (eastern Sichuan), as well as the priority areas revealed under RCP2.6. The areas at a high-risk of losing weighted endemism were almost equivalent to those for species richness under RCP2.6 and RCP 8.5, but the ratio covered by existing nature reserves are higher (varying from 7.1% to 11.2%) than for species richness. Our prediction showed that Zhangzhou (southern Fujian) Huizhou,

Fig. 7. Distribution of change in species richness for increasing elevation (m) under climate change alone [(a) and (c)] and under combined climate and land use

change [(b) and (d)] of RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’perfect-dispersal’ scenario.

(10)

Shenzhen (central Guangdong), and south Taiwan would tend to lose many rare and common species under the climate change only scenario, just as in the locations we mentioned above.

4. Discussion

4.1. Divergence of species response to future climate and land use change

Our results showed that generally 52% of the Chinese endemic

Rhododendron species would expand their geographic range, while 48%

of the species would contract their geographic range under different climate and land use change of ’perfect-dispersal’ scenarios. However, species showed varying spatial changes when considering different greenhouse gas emissions, and change ratios were generally higher under RCP8.5 than under RCP2.6. To some extent, this result of equal losses and gains was unexpected in light of earlier evidence from Europe (Thuiller et al., 2011, 2005) and North America (McKenney et al., 2007; Zhang et al., 2017), which predicted that the majority of plant species

would experience large contractions in response to climate change. As well, land use change was expected to aggravate species loss (Di Marco et al., 2019; Sirami et al., 2017; Titeux et al., 2016).

However, a number of studies did show that many species may also ’benefit’ from future climate change, deeming it likely that they expand their habitats. Forest herbs (Skov and Svenning, 2004), beech in Europe (Harrison et al., 2006), fungus species in the Nepali Himalaya (Shrestha and Bawa, 2014) and shrubs in the Arctic, high-latitude and alpine tundra ecosystems (Myers-Smith et al., 2011) forming examples of this. Cannone et al. (2007) found that shrubs showed rapid expansion rates of 5.6% per decade at elevations between 2400 and 2500 m in site of Eu-ropean Alps. In our case, we speculate that the relatively high proportion of Rhododendron species that experience gains in suitable areas rather than losses is caused by Rhododendron being a typical alpine and sub-alpine genus of mostly shrubs. Therefore, they fit the profile of species that have been identified in earlier studies to benefit from the shortening of the snow cover season (Keller et al. 2005). A 1–2 ◦_{C warming of the air} temperature combined with increasing summer precipitation may result

Fig. 8. Distribution of change in species richness for increasing elevation (m) under climate change alone [(a) and (c)] and under combined climate and land use

change [(b) and (d)] of RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’no-dispersal’ scenario.

(11)

in the expansion of plant species (Cannone et al., 2007). Therefore, we suggest that species may respond to future environmental change indi-vidually (Wang et al., 2019; Yu et al., 2019), but climate and land use change caused by differences in greenhouse gas emissions will have great influence on the diversity pattern of Rhododendrons, making them valuable indicators of environmental change for alpine and subalpine ecosystems.

By comparison, about 7% (ranging 5%~9%) of Rhododendron spe-cies would keep ’still’ while the remaining spespe-cies would lose habitats to a varying degree under different future climate and land use of ’no- dispersal’ scenario. The relatively pessimistic predictions may change if a more realistic dispersal scenario incorporated (Feeley and Silman, 2010). How species respond to future environmental change in reality is a complicated issue which would be influenced by genetic variation (evolution), disturbance (including climate and land-use change), dispersal ability (Travis et al., 2013), and interaction of all the factors

(Harmon et al., 2009). Therefore, we emphasize the importance of including sound variables (scenarios, e.g., dispersal ability) or applying a species-specific model for predicting species distribution under further environmental change, especially aiming to make conservation recommendations.

4.2. Higher extinction risk at lower elevations

A relatively low proportion (48%) of Rhododendron species was predicted to lose habitat in the future. However, we did find that the decrease in species richness and weighted endemism of Chinese endemic

Rhododendron species would be more severe under combined climate

and land use change than under climate change alone for both the emission (i.e., RCP2.6 and the RCP8.5) and dispersal (i.e., ’perfect-’ and ’no-dispersal’) scenario. This result is in line with previous findings that the joint effect of climate change and land use change on species

Fig. 9. Spatial patterns of change in weighted endemism under climate change alone [(a) and (c)] and under combined climate and land use change [(b) and (d)] for

the RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’perfect-dispersal’ scenario for the 2070s (2071–2100).

(12)

distribution and biodiversity is more pronounced, leading to a greater risk of species extinction, than when climate change in considered on its own (Frishkoff et al., 2016; Oliver and Morecroft, 2014). We therefore emphasize that land use change remains a threat to biodiversity (Ost-berg et al., 2015), also in China, and should be included as an important factor when predicting future distribution of species and changes in diversity patterns (Newbold et al., 2018).

An important result of this study is the prediction that the loss of species richness and weighted endemism will primarily occur at lower elevations, while gains are predicted to occur at higher elevations. This result is also consistent with previous studies. Oliver and Morecroft (2014) concluded that species richness of Californian butterflies had declined at lower elevations where land use was more intensive. And a recent study concluded that higher species vulnerability occurred at the lower elevations and mountain bases as a result of intense human

pressures globally (Elsen et al., 2020). We infer that for Rhododendron species that experience contraction of suitable habitat at low elevation, this contraction would mainly stem from the effect of land use change. Human activity, including road construction, agricultural encroach-ment, and expansion of the urban environencroach-ment, has largely been occurring at low elevation (Ma et al., 2014), causing habitat loss as has been demonstrated worldwide (Feeley and Silman, 2010). Hence, we emphasize the importance of protecting species at lower elevation, and especially areas of southeastern China.

4.3. Conservation priority areas

As one of the largest and oldest genera in China (Ma et al., 2014; MacKay and Gardiner, 2016), the genus Rhododendron, with its complex taxonomic structure and morphological variation, is an excellent

Fig. 10. Spatial patterns of change in weighted endemism under climate change alone [(a) and (c)] and under combined climate and land use change [(b) and (d)]

for the RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’no-dispersal’ scenario for the 2070s (2071–2100).

(13)

candidate for conservation planning (MacKay and Gardiner, 2016). Our results show that spatially representative and complementary sites for

Rhododendron diversity are poorly covered by the existing protected

areas, and that a fairly large proportion (>89%) of protected areas is not designed to efficiently protect Rhododendron species in their natural habitat on a national scale. Projected future gaps in conservation areas concerning Chinese endemic Rhododendron species were identified using species richness and weighted endemism. Large gap areas in the existing protected area network pose a great challenge for biodiversity conser-vation. This is consistent with a previous study by Xu et al. (2017), which showed that current Chinese protected areas only cover 13% of threatened plant species and are not able to provide the long-term protection of biodiversity and ecosystem services necessary to achieve sustainable development. The areas prioritized regarding the need for

protection include the parallel ridge-and-valley area of eastern Sichuan, southeastern Tibet, western Yunnan, the southern part of Shaanxi, and scattered areas of Guangdong, Hainan and Taiwan.

It is worth noting that the change in species richness and weighted endemism in Sichuan can be attributed to its unique topography and climate. Western Sichuan consists of numerous mountain ranges (e.g., the Daxue Mountains and Gongga Shan) forming the easternmost part of the Tibetan Plateau (Wu et al., 2017), which are able to provide more suitable habitat areas for Rhododendrons under environmental change (Liang et al., 2018). However, eastern Sichuan is mostly within the fertile Sichuan basin, which is one of the most densely populated regions of China influenced by rapid urbanization as well as a statistically sig-nificant warming trend (approximately 4 ◦_{C by the end of the twenty-} first century), with the greatest warming occurring over the central

Fig. 11. Distribution of change in weighted endemism for increasing elevation under climate change alone [(a) and (c)] and under combined climate and land use

change [(b) and (d)] of RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’perfect-dispersal’ scenario.

(14)

Fig. 12. Distribution of change in weighted endemism for increasing elevation under climate change alone [(a) and (c)] and under combined climate and land use

change [(b) and (d)] of RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] of ’no-dispersal’ scenario.

Table 1

High-risk areas and areas covered by existing nature reserves under climate change as well as the combination of climate and land use change of both ’perfect-’ and ’no- dispersal’ scenarios for the 2070s (2071–2100).

Indices Scenarios Total of high-risk areas (km2) _{Gap areas (km}2 _{Areas covered by nature reserves (km}2₎ _{Percentage covered by nature reserve (%)}

Species richness RCP2.6 CC 16,666 16,600 66 0.4 CCLC 21,800 21,620 180 0.8 CC 18,360 17,500 860 4.7 RCP8.5 CCLC 28,000 26,200 1800 6.4 Weighted endemism RCP2.6 CC CCLC 13,300 14,000 12,000 13,000 1300 1000 9.8 7.1 RCP8.5 CC 13,100 12,000 1100 8.3 CCLC 16,900 15,000 1900 11.2

CC: Climate Change only. CCLC: Climate Change and Land use Change.

(15)

plains of the Sichuan basin (Bannister et al., 2017). Tang et al. (2006) showed that the Sichuan basin is one of the hotspots regarding endan-gered plants in China. And Tang et al. (2017) concluded that new nature reserves need to be established urgently in the mountainous margins of the Sichuan basin to protect a tertiary relict tree species threatened under future climate change. Lu et al. (2020) also concluded that Sichuan is one of the biodiversity hotspots most threatened by climate warming and geological disasters, soil erosion, and agricultural recla-mation. We therefore propose that protected areas in China should include more mountainous terrain as well as part of southern China, which is experiencing dramatic land use changes, to ensure future environmental change can be better withstood.

4.4. Uncertainty

Although our study addressed some shortcomings of previous studies, there are still a number of uncertainties that would benefit from more thorough investigation. Firstly, although we used the most com-plete Rhododendron genus database available to model the endemic species as a representative subset of all plant species in China, having a more complete coverage of all flora at country level would allow for greater insight into the spatial structure and to sensitivity to changes regarding China’s flora, aiding the ability to highlight priority areas of species richness and endemism. Secondly, there are distribution model algorithm uncertainties. Many different algorithms exist to model the distributions of species (Wiens et al., 2009). Since our species database contains only presence data, using MaxEnt was an ideal choice, but every static modelling approach has its limitations and uncertainties

Fig. 13. Predicted priority areas for the conservation of Chinese endemic Rhododendron species determined by species richness under climate change only [(a) and

(c)] and under combined climate and land use change [(b) and (d)] for the RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] scenarios in China for the 2070 s (2071–2100). ND: no-dispersal; PD: perfect-dispersal.

(16)

(Franklin, 2010b). Last but not least, our models assumed all species are affected only by climate and land use change, combining two potential extreme dispersal (i.e., ’perfect-dispersal’ and ’no-dispersal’) scenarios. This is a simplification of the real effects of environmental change and species growth. And a more realistic dispersal scenario, or the other variables playing a role in the process of establishment and growth of

Rhododendron were not included in the current analysis.

5. Conclusions

In this study, we estimated the potential synergistic impacts of future climate and land use change on 191 Chinese endemic Rhododendron species. We found that 52% of endemic Rhododendron species were predicted to expand and shift their geographic ranges, typically to the northwest and north, under the future emission (i.e., RCP2.6 and

RCP8.5) and ’perfect-dispersal’ scenarios. The remaining 48% of

Rhododendron species were predicted to contract in geographic range,

particularly at lower elevations. Meanwhile, about 7% of Rhododendron would keep ’still’ and the rest of species would have shirking habitats under the ’no-dispersal’ scenario. We established that the existing pro-tected areas only cover a very small portion (0.4%~11.2%) of the areas at high-risk of losing Rhododendron under the varying future environ-mental change scenarios. Based on these results, we conclude that Chi-nese endemic Rhododendron species at lower elevations are highly vulnerable to climate and land use change and face an elevated risk of extinction, thereby calling for more attention and protection. We high-light the critical role of endemic Rhododendron species as good in-dicators for measuring, evaluating and understanding the effectiveness of our biodiversity conservation efforts. Our work provide insight into the status, trends and threats for the Chinese endemic Rhododendron

Fig. 14. Predicted priority areas for the conservation of Chinese endemic Rhododendron species determined by weighted endemism under climate change only [(a)

and (c)] and combined climate and land use change [(b) and (d)] for the RCP2.6 [(a) and (b)] and RCP8.5 [(c) and (d)] scenarios in China for the 2070s (2071–2100). ND: no-dispersal; PD: perfect-dispersal.

(17)

Ecological Indicators 126 (2021) 107699 species, presenting potentially valuable indicators of climate and land

use change for alpine and subalpine ecosystems, and proposing large areas that require prioritized attention for plant biodiversity conserva-tion in a rapidly changing world.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 41901060 and 41671430). We are grateful to Dr. Wenyun Zuo for her data-sharing initiative.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ecolind.2021.107699.

References

Aguirre-Gutierrez, J., Serna-Chavez, H.M., Villalobos-Arambula, A.R., Perez de la Rosa, J.A., Raes, N., 2015. Similar but not equivalent: ecological niche comparison across closely-related Mexican white pines. Divers. Distrib. 21, 245–257.

Alagador, D., Cerdeira, J.O., Araújo, M.B., Saura, S., 2014. Shifting protected areas: scheduling spatial priorities under climate change. J. Appl. Ecol. 51, 703–713.

Allouche, O., Tsoar, A., Kadmon, R., 2006. Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS). J. Appl. Ecol. 43, 1223–1232.

Bannister, D., Herzog, M., Graf, H.-F., Hosking, J.S., Short, C.A., 2017. An assessment of recent and future temperature change over the Sichuan Basin, China, Using CMIP5 Climate Models. J. Clim. 30, 6701–6722.

Brook, B.W., Sodhi, N.S., Bradshaw, C.J.A., 2008. Synergies among extinction drivers under global change. Trends Ecol. Evol. 23, 453–460.

Brown, J.L., Bennett, J.R., French, C.M., 2017. SDMtoolbox 2.0: the next generation Python-based GIS toolkit for landscape genetic, biogeographic and species distribution model analyses. PeerJ 5, e4095.

Cannone, N., Sgorbati, S., Guglielmin, M., 2007. Unexpected impacts of climate change on alpine vegetation. Front. Ecol. Environ. 5, 360–364.

Chen, Y., Zhang, J., Jiang, J., Nielsen, S.E., He, F., Robertson, M., 2017. Assessing the effectiveness of China’s protected areas to conserve current and future amphibian diversity. Divers. Distrib. 23, 146–157.

Di Marco, M., Harwood, T.D., Hoskins, A.J., Ware, C., Hill, S.L.L., Ferrier, S., 2019. Projecting impacts of global climate and land-use scenarios on plant biodiversity using compositional-turnover modelling. Glob. Change Biol. 25, 2763–2778.

Distler, T., Schuetz, J.G., Velasquez-Tibata, J., Langham, G.M., 2015. Stacked species distribution models and macroecological models provide congruent projections of avian species richness under climate change. J. Biogeogr. 42, 976–988.

Elsen, P.R., Monahan, W.B., Merenlender, A.M., 2020. Topography and human pressure in mountain ranges alter expected species responses to climate change. Nat. Commun. 11, 1974.

ESRI, 2011. ArcGIS Desktop: Release 10.2. Environmental Systems Research Institute (ESRI), Redlands, CA, U.S.A.

Feeley, K.J., Silman, M.R., 2010. Land-use and climate change effects on population size and extinction risk of Andean plants. Glob. Change Biol. 16, 3215–3222.

Franklin, J., 2010a. Mapping Species Distributions: Spatial Inference and Prediction. Cambridge University Press.

Franklin, J., 2010b. Moving beyond static species distribution models in support of conservation biogeography. Divers. Distrib. 16, 321–330.

Frishkoff, L.O., Karp, D.S., Flanders, J.R., Zook, J., Hadly, E.A., Daily, G.C., M’Gonigle, L. K., 2016. Climate change and habitat conversion favour the same species. Ecol. Lett. 19, 1081–1090.

Gibbs, D., Chamberlain, D., Argent, G., 2011. The Red List of Rhododendrons. Botanic Gardens Conservation International, Richmond, UK.

Harmon, J.P., Moran, N.A., Ives, A.R., 2009. Species response to environmental change: impacts of food web interactions and evolution. Science 323, 1347.

Harrison, P.A., Berry, P.M., Butt, N., New, M., 2006. Modelling climate change impacts on species’ distributions at the European scale: implications for conservation policy. Environ. Sci. Policy 9, 116–128.

Hengl, T., de Jesus, J.M., MacMillan, R.A., Batjes, N.H., Heuvelink, G.B., Ribeiro, E., Samuel-Rosa, A., Kempen, B., Leenaars, J.G., Walsh, M.G., Gonzalez, M.R., 2014. SoilGrids1km–global soil information based on automated mapping. PLoS One 9, e105992.

Herkt, K.M.B., Barnikel, G., Skidmore, A.K., Fahr, J., 2016. A high-resolution model of bat diversity and endemism for continental Africa. Ecol. Model. 320, 9–28.

IPCC, 2013. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Kumar, P., 2012. Assessment of impact of climate change on Rhododendrons in Sikkim Himalayas using Maxent modelling: limitations and challenges. Biodivers. Conserv. 21, 1251–1266.

Laffan, S.W., Crisp, M.D., 2003. Assessing endemism at multiple spatial scales, with an example from the Australian vascular flora. J. Biogeogr. 30, 511–520.

Li, X., Yu, L., Sohl, T., Clinton, N., Li, W., Zhu, Z., Liu, X., Gong, P., 2016. A cellular automata downscaling based 1 km global land use datasets (2010–2100). Sci. Bull. 61, 1651–1661.

Liang, Q., Xu, X., Mao, K., Wang, M., Wang, K., Xi, Z., Liu, J., 2018. Shifts in plant distributions in response to climate warming in a biodiversity hotspot, the Hengduan Mountains. J. Biogeogr. 45, 1334–1344.

Liu, C., White, M., Newell, G., Pearson, R., 2013. Selecting thresholds for the prediction of species occurrence with presence-only data. J. Biogeogr. 40, 778–789.

Lu, Y., Yang, Y., Sun, B., Yuan, J., Yu, M., Stenseth, N.C., Bullock, J.M., Obersteiner, M., 2020. Spatial variation in biodiversity loss across China under multiple environmental stressors. Sci. Adv. 6, eabd0952.

Ma, Y., Nielsen, J., Chamberlain, D.F., Li, X., Sun, W., 2014. The conservation of

Rhododendrons is of greater urgency than has been previously acknowledged in

China. Biodivers. Conserv. 23, 3149–3154.

MacKay, M., Gardiner, S.E., 2016. A model for determining ex situ conservation priorities in big genera is provided by analysis of the subgenera of Rhododendron (Ericaceae). Biodivers. Conserv. 26, 189–208.

McKenney, D.W., Pedlar, J.H., Lawrence, K., Campbell, K., Hutchinson, M.F., 2007. Potential impacts of climate change on the distribution of North American trees. Bioscience 57, 939–948.

Merow, C., Smith, M.J., Edwards Jr., T.C., Guisan, A., McMahon, S.M., Normand, S., Thuiller, W., Wueest, R.O., Zimmermann, N.E., Elith, J., 2014. What do we gain from simplicity versus complexity in species distribution models? Ecography 37, 1267–1281.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853–858.

Myers-Smith, I.H., Forbes, B.C., Wilmking, M., Hallinger, M., Lantz, T., Blok, D., Tape, K. D., Macias-Fauria, M., Sass-Klaassen, U., L´evesque, E., Boudreau, S., Ropars, P., Hermanutz, L., Trant, A., Collier, L.S., Weijers, S., Rozema, J., Rayback, S.A., Schmidt, N.M., Schaepman-Strub, G., Wipf, S., Rixen, C., M´enard, C.B., Venn, S., Goetz, S., Andreu-Hayles, L., Elmendorf, S., Ravolainen, V., Welker, J., Grogan, P., Epstein, H.E., Hik, D.S., 2011. Shrub expansion in tundra ecosystems: dynamics, impacts and research priorities. Environ. Res. Lett. 6, 045509.

Newbold, T., Hudson, L.N., Contu, S., Hill, S.L.L., Beck, J., Liu, Y., Meyer, C., Phillips, H. R.P., Scharlemann, J.P.W., Purvis, A., 2018. Widespread winners and narrow-ranged losers: land use homogenizes biodiversity in local assemblages worldwide. PLoS Biol. 16, e2006841.

Oliver, T.H., Morecroft, M.D., 2014. Interactions between climate change and land use change on biodiversity: attribution problems, risks, and opportunities. Wiley Interdiscip. Rev. Clim. Change 5, 317–335.

Osborn, D., Newbold, T., Adams, G.L., Albaladejo Robles, G., Boakes, E.H., Braga Ferreira, G., Chapman, A.S.A., Etard, A., Gibb, R., Millard, J., Outhwaite, C.L., Williams, J.J., 2019. Climate and land-use change homogenise terrestrial biodiversity, with consequences for ecosystem functioning and human well-being. Emerging Topics Life Sci. 3, 207–219.

Ostberg, S., Schaphoff, S., Lucht, W., Gerten, D., 2015. Three centuries of dual pressure from land use and climate change on the biosphere. Environ. Res. Lett. 10, 044011.

Pearson, R.G., Raxworthy, C.J., Nakamura, M., Peterson, A.T., 2007. Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. J. Biogeogr. 34, 102–117.

Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecol. Model. 190, 231–259.

R Core Team, 2015. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project. org/.

Rana, S.K., Rana, H.K., Luo, D., Sun, H., 2021. Estimating climate-induced ‘Nowhere to go’ range shifts of the Himalayan Incarvillea Juss. using multi-model median ensemble species distribution models. Ecol. Ind. 121, 107127.

Sala, O.E., Chapin, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber- Sanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M., Wall, D.H., 2000. Biodiversity - global biodiversity scenarios for the year 2100. Science 287, 1770–1774.

Shrestha, U.B., Bawa, K.S., 2014. Impact of climate change on potential distribution of Chinese caterpillar fungus (Ophiocordyceps sinensis) in Nepal Himalaya. PLoS One 9.

Sirami, C., Caplat, P., Popy, S., Clamens, A., Arlettaz, R., Jiguet, F., Brotons, L., Martin, J.-L., 2017. Impacts of global change on species distributions: obstacles and solutions to integrate climate and land use. Glob. Ecol. Biogeogr. 26, 385–394.

Skov, F., Svenning, J.C., 2004. Potential impact of climatic change on the distribution of forest herbs in Europe. Ecography 27, 366–380.

Sommer, J.H., Kreft, H., Kier, G., Jetz, W., Mutke, J., Barthlott, W., 2010. Projected impacts of climate change on regional capacities for global plant species richness. Proc. R. Soc. B: Biol. Sci. 277, 2271–2280.

Swets, J.A., 1988. Measuring the accuracy of diagnostic systems. Science 240, 1285–1293.

(18)

Ecological Indicators 126 (2021) 107699 Synge, H., London, L.S.o., Isles, B.S.o.t.B., 1981. The Biological Aspects of Rare Plant

Conservation. Wiley.

Tang, C.Q., Dong, Y.F., Herrando-Moraira, S., Matsui, T., Ohashi, H., He, L.Y., Nakao, K., Tanaka, N., Tomita, M., Li, X.S., Yan, H.Z., Peng, M.C., Hu, J., Yang, R.H., Li, W.J., Yan, K., Hou, X., Zhang, Z.Y., Lopez-Pujol, J., 2017. Potential effects of climate change on geographic distribution of the Tertiary relict tree species Davidia involucrata in China. Sci. Rep. 7, 43822.

Tang, Z., Wang, Z., Zheng, C., Fang, J., 2006. Biodiversity in China’s mountains. Front. Ecol. Environ. 4, 347–352.

ter Steege, H., Pitman, N.C.A., Sabatier, D., Baraloto, C., Salomão, R.P., Guevara, J.E., Phillips, O.L., Castilho, C.V., Magnusson, W.E., Molino, J.-F., Monteagudo, A., Núñez Vargas, P., Montero, J.C., Feldpausch, T.R., Coronado, E.N.H., Killeen, T.J., Mostacedo, B., Vasquez, R., Assis, R.L., Terborgh, J., Wittmann, F., Andrade, A., Laurance, W.F., Laurance, S.G.W., Marimon, B.S., Marimon, B.-H., Guimarães Vieira, I.C., Amaral, I.L., Brienen, R., Castellanos, H., Cárdenas López, D., Duivenvoorden, J. F., Mogollón, H.F., Matos, F.D.d.A., Dávila, N., García-Villacorta, R., Stevenson Diaz, P.R., Costa, F., Emilio, T., Levis, C., Schietti, J., Souza, P., Alonso, A., Dallmeier, F., Montoya, A.J.D., Fernandez Piedade, M.T., Araujo-Murakami, A., Arroyo, L., Gribel, R., Fine, P.V.A., Peres, C.A., Toledo, M., Aymard C., G.A., Baker, T.R., Cerón, C., Engel, J., Henkel, T.W., Maas, P., Petronelli, P., Stropp, J., Zartman, C.E., Daly, D., Neill, D., Silveira, M., Paredes, M.R., Chave, J., Lima Filho, D.d.A., Jørgensen, P.M., Fuentes, A., Schöngart, J., Cornejo Valverde, F., Di Fiore, A., Jimenez, E.M., Peñuela Mora, M.C., Phillips, J.F., Rivas, G., van Andel, T.R., von Hildebrand, P., Hoffman, B., Zent, E.L., Malhi, Y., Prieto, A., Rudas, A., Ruschell, A.R., Silva, N., Vos, V., Zent, S., Oliveira, A.A., Schutz, A.C., Gonzales, T., Trindade Nascimento, M., Ramirez- Angulo, H., Sierra, R., Tirado, M., Umaña Medina, M.N., van der Heijden, G., Vela, C. I.A., Vilanova Torre, E., Vriesendorp, C., Wang, O., Young, K.R., Baider, C., Balslev, H., Ferreira, C., Mesones, I., Torres-Lezama, A., Urrego Giraldo, L.E., Zagt, R., Alexiades, M.N., Hernandez, L., Huamantupa-Chuquimaco, I., Milliken, W., Palacios Cuenca, W., Pauletto, D., Valderrama Sandoval, E., Valenzuela Gamarra, L., Dexter, K.G., Feeley, K., Lopez-Gonzalez, G., Silman, M.R., 2013. Hyperdominance in the Amazonian Tree Flora. Science 342, 1243092.

Thomas, C.D., Cameron, A., Green, R.E., Bakkenes, M., Beaumont, L.J., Collingham, Y.C., Erasmus, B.F.N., De Siqueira, M.F., Grainger, A., Hannah, L., Hughes, L., Huntley, B., Van Jaarsveld, A.S., Midgley, G.F., Miles, L., Ortega-Huerta, M.A., Peterson, A.T., Phillips, O.L., Williams, S.E., 2004. Extinction risk from climate change. Nature 427, 145–148.

Thuiller, W., Lavorel, S., Araujo, M.B., Sykes, M.T., Prentice, I.C., 2005. Climate change threats to plant diversity in Europe. PNAS 102, 8245–8250.

Thuiller, W., Lavergne, S., Roquet, C., Boulangeat, I., Lafourcade, B., Araujo, M.B., 2011. Consequences of climate change on the tree of life in Europe. Nature 470, 531–534.

Titeux, N., Henle, K., Mihoub, J.B., Regos, A., Geijzendorffer, I.R., Cramer, W., Verburg, P.H., Brotons, L., 2016. Biodiversity scenarios neglect future land-use changes. Glob. Change Biol. 22, 2505–2515.

Travis, J.M.J., Delgado, M., Bocedi, G., Baguette, M., Barto´n, K., Bonte, D., Boulangeat, I., Hodgson, J.A., Kubisch, A., Penteriani, V., Saastamoinen, M., Stevens, V.M., Bullock, J.M., 2013. Dispersal and species’ responses to climate change. Oikos 122, 1532–1540.

van Proosdij, A.S.J., Sosef, M.S.M., Wieringa, J.J., Raes, N., 2016. Minimum required number of specimen records to develop accurate species distribution models. Ecography 39, 542–552.

van Vuuren, D.P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G.C., Kram, T., Krey, V., Lamarque, J.-F., Masui, T., Meinshausen, M., Nakicenovic, N., Smith, S.J., Rose, S.K., 2011. The representative concentration pathways: an overview. Clim. Change 109, 5–31.

Wang, J., Wang, Y., Feng, J., Chen, C., Chen, J., Long, T., Li, J., Zang, R., Li, J., 2019. Differential responses to climate and land-use changes in threatened Chinese Taxus Species. Forests 10, 766.

Wiens, J.A., Stralberg, D., Jongsomjit, D., Howell, C.A., Snyder, M.A., 2009. Niches, models, and climate change: assessing the assumptions and uncertainties. PNAS 106, 19729–19736.

Wu, Y., DuBay, S.G., Colwell, R.K., Ran, J., Lei, F., 2017. Mobile hotspots and refugia of avian diversity in the mountains of south-west China under past and contemporary global climate change. J. Biogeogr. 44, 615–626.

Wu, Z.Y., Peter, H.R., Hong, D.Y., 2005. Flora of China. Science Press, Beijing.

Xu, H., Cao, M., Wang, Z., Wu, Y., Cao, Y., Wu, J., Le, Z., Cui, P., Ding, H., Xu, W., Peng, H., Jiang, J., Wu, Y., Jiang, X., Zhang, Z., Rao, D., Li, J., Lei, F., Xia, N., Han, L., Cao, W., Wu, J., Xia, X., Li, Y., 2018. Low ecological representation in the protected area network of China. Ecol. Evol. 8, 6290–6298.

Xu, W., Xiao, Y., Zhang, J., Yang, W., Zhang, L., Hull, V., Wang, Z., Zheng, H., Liu, J., Polasky, S., Jiang, L., Xiao, Y., Shi, X., Rao, E., Lu, F., Wang, X., Daily, G.C., Ouyang, Z., 2017. Strengthening protected areas for biodiversity and ecosystem services in China. Proc. Natl. Acad. Sci. 114, 1601–1606.

Ye, X., Yu, X., Yu, C., Tayibazhaer, A., Xu, F., Skidmore, A.K., Wang, T., 2018. Impacts of future climate and land cover changes on threatened mammals in the semi-arid Chinese Altai Mountains. Sci. Total Environ. 612, 775–787.

Yu, F.Y., Wang, T.J., Groen, T.A., Skidmore, A.K., Yang, X.F., Geng, Y.Y., Ma, K.P., 2015. Multi-scale comparison of topographic complexity indices in relation to plant species richness. Ecol. Complexity 22, 93–101.

Yu, F.Y., Skidmore, A.K., Wang, T.J., Huang, J.H., Ma, K.P., Groen, T.A., 2017.

Rhododendron diversity patterns and priority conservation areas in China. Divers.

Distrib. 23, 1143–1156.

Yu, F.Y., Wang, T.J., Groen, T.A., Skidmore, A.K., Yang, X.F., Ma, K.P., Wu, Z.F., 2019. Climate and land use changes will degrade the distribution of Rhododendrons in China. Sci. Total Environ. 659, 515–528.

Zhang, J., Nielsen, S.E., Chen, Y., Georges, D., Qin, Y., Wang, S.-S., Svenning, J.-C., Thuiller, W., 2017. Extinction risk of North American seed plants elevated by climate and land-use change. J. Appl. Ecol. 54, 303–312.